AQL Sample Size for Battery Inspection

[terranceyu]

Hello everyone,

I have one concern on the AQL sampling on battery inspection.

Currently production found one battery failed FCT test - low voltage. Production quantity : 500 pcs.

IQC is measuring the battery voltage using multimeter. Sample size : 4 pcs. Lot quantity: 5000 pcs. Are we suppose to use the AQL sampling size according to ISO 2859-1, to respond to this complaint? Or keep at actual practice, while sending the faulty part to supplier for investigation.

Please, share your thought on this. Thanks.

Terrance

 

[reynald]

If the sample size was special due to volume constraints, you can use probability theory to test the hypothesis that the %defective increased in reference to the historical baseline.

My logic would go as follows:

Decision:
if %d is <= some value--> ignore the problem
if %d is > some value--> highlight the problem

Use a p-test
Ho: p is same as hostorical level
Ha: P incresed from historical level

then proceed to whatever conlusion the p-test takes you.

 

[terranceyu]

If the sample size was special due to volume constraints, you can use probability theory to test the hypothesis that the %defective increased in reference to the historical baseline.

My logic would go as follows:

Decision:
if %d is <= some value--> ignore the problem
if %d is > some value--> highlight the problem

Use a p-test
Ho: p is same as hostorical level
Ha: P incresed from historical level

then proceed to whatever conlusion the p-test takes you.

In the sametime, I also using the same theory feedback to him. But he insisted to check the battery lot exactly according to ISO 2859-1. Does this actually serve any purpose?

What if the battery checked as per ISO 2859-1 sample size, so can we judge this lot as good, but in the end of line we found one piece defect after 100% FCT test?

 

[terranceyu]

In the sametime, I also using the same theory feedback to him. But he insisted to check the battery lot exactly according to ISO 2859-1. Does this actually serve any purpose?

What if the battery checked as per ISO 2859-1 sample size, so can we judge this lot as good, but in the end of line we found one piece defect after 100% FCT test?

Any advice for me, dear covers?

 

[Bev D]

I am not sure what you are asking?

Please clarify the relationship of yourself, IQC, the test location and 'he'. who is 'he' that is complaining? (Are you in IQC and 'he' is your manufacturing line?)

What is the ISO2859-1 sample size?
If different from n=4 why did you choose n=4?

It is quite possible that using a different sample size will give you a better understanding of the quality of the lot.

 

[terranceyu]

I am not sure what you are asking?

Please clarify the relationship of yourself, IQC, the test location and 'he'. who is 'he' that is complaining? (Are you in IQC and 'he' is your manufacturing line?)

What is the ISO2859-1 sample size?
If different from n=4 why did you choose n=4?

It is quite possible that using a different sample size will give you a better understanding of the quality of the lot.

I am IQC engineer. He is the project engineer which is very much concerned for this medical product, if the battery voltage going below requirement, and uncovered by Functional test, the user might not able to get the sufficient dosage due to insuffient energy source.

ISO2859-1 sample size is 200 pcs. I am choosing 4 pcs for voltage measurement due to this is a product that has been checked by supplier 100%, and lack of human resources to actually doing this, because we have almost 6000 items to be checked in one single day!

IMO, we should take care of the battery storage condition, shelf life, and get supplier corrective action and preventive action, other than just depending on AQL inspection at receiving. It should minimize the risk, and satisfied everybody.

Pls, let me know your thoughts. Good advices are welcomed.

 

[terranceyu]

Hi Bev D, any advices?

 

[Bev D]

I am IQC engineer. He is the project engineer which is very much concerned for this medical product, if the battery voltage going below requirement, and uncovered by Functional test, the user might not able to get the sufficient dosage due to insuffient energy source.

ISO2859-1 sample size is 200 pcs. I am choosing 4 pcs for voltage measurement due to this is a product that has been checked by supplier 100%, and lack of human resources to actually doing this, because we have almost 6000 items to be checked in one single day!

IMO, we should take care of the battery storage condition, shelf life, and get supplier corrective action and preventive action, other than just depending on AQL inspection at receiving. It should minimize the risk, and satisfied everybody.

Pls, let me know your thoughts. Good advices are welcomed.

The process engineer is correct to be concerned and your lack of human resources to perform incoming inspection will not be an excuse if there is a field action on the instrument due to battery voltage.

You are correct that storage and shelf life should be controlled and the supplier should recieve a corrective action.

However, sampling only 4 batteries tells you nothing about the lot and is a waste of time. There are two plans that can be used: the one you mention is for pass/fail data (the battery meets specification or it doesn't) and these will always have larger sample sizes. A second option is to use sampling for continuous data (or as some call it "variable" data). This will get your sample sizes down but wstill allow you to perform a meaningful inspection.

 

[terranceyu]

The process engineer is correct to be concerned and your lack of human resources to perform incoming inspection will not be an excuse if there is a field action on the instrument due to battery voltage.

You are correct that storage and shelf life should be controlled and the supplier should recieve a corrective action.

However, sampling only 4 batteries tells you nothing about the lot and is a waste of time. There are two plans that can be used: the one you mention is for pass/fail data (the battery meets specification or it doesn't) and these will always have larger sample sizes. A second option is to use sampling for continuous data (or as some call it "variable" data). This will get your sample sizes down but wstill allow you to perform a meaningful inspection.

Hi Bev D,

Can you provide in details how sampling for continuous data is done. How many samples should be taken for this? Based on AQL level or....?

Your help is very much appreciated.

 

[Jim Wynne]

However, sampling only 4 batteries tells you nothing about the lot and is a waste of time.

I agree with this in spirit but I think clarification might be good. If the OP discovered a single defective in a 4-piece sample from a 500-piece lot, the effort wasn't a waste of time. What is wrong, as you point out, is that a sample size of 4 isn't reasonable given the lot size and the risk involved.

If this is a situation where one defective is too many and the risk is high, it's possible that sampling inspection is inappropriate altogether.